Introduction
Welcome to the technical guidelines of the Swiss Digital Pathology Algorithm Alliance (SDiPAA)! We are a collaborative network of Swiss organizations, established with the mission of facilitating the sharing, deployment, and optimization of digital pathology algorithms across Switzerland. Our collective efforts aim to advance the field of digital pathology through standardized practices that ensure interoperability, reproducibility, and high performance across diverse computational environments.
As digital pathology increasingly relies on computational tools for tasks such as image analysis, diagnosis, and prognostication, the need for well-structured, maintainable, and efficient code has become paramount. These guidelines have been developed to provide a comprehensive framework for structuring inference code, tailored specifically for the needs of our community.
About These Guidelines
The guidelines outlined in this document encompass several critical areas necessary for the successful deployment and sharing of digital pathology algorithms:
- Code Structure Recommendations: We provide best practices for organizing your projects and illustrate these concepts using python to enhance clarity, maintainability, and collaboration.
- Containerization of Code: To ensure consistency across different environments, we advocate for the containerization of code using tools like Docker. This section covers the creation of portable and reproducible environments that can be easily shared and deployed across institutions.
- Validation and Benchmarking Algorithms: Accurate benchmarking is crucial for assessing the performance of algorithms in real-world scenarios as well as identifying weakness in the algorithm. We stress the importance of documenting all aspects of the computational environment ensuring that others can replicate your results, assess the qualities and weaknesses of algorithms and estimate inference time with precision.
Our Commitment
As a part of the Swiss Digital Pathology Algorithm Alliance, we are committed to fostering innovation through collaboration and knowledge sharing. By adhering to these guidelines, we aim to ensure that the digital pathology algorithms developed within our network are robust, scalable, and easily transferable between organizations. We encourage all contributors to familiarize themselves with these guidelines. Together, we can drive forward the development of cutting-edge digital pathology solutions that benefit the entire healthcare community.
Document Downloads
For additional information, please contact the working group leader: Petros Liakopoulos Petros.Liakopoulos@unige.ch